image_8295e-COVID-19.jpg

In [1]:
#path = 'https://raw.githubusercontent.com/umangkejriwal1122/Machine-Learning/master/Data%20Sets/covid_19_clean_complete.csv'
In [2]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns 
import plotly.express as px
In [3]:
df = pd.read_csv('C:/Users/VAIO/Downloads/COVID-19 AI ML/covid_19_clean_complete.csv')
df.head()
Out[3]:
Province/State Country/Region Lat Long Date Confirmed Deaths Recovered
0 NaN Afghanistan 33.0000 65.0000 1/22/20 0 0 0
1 NaN Albania 41.1533 20.1683 1/22/20 0 0 0
2 NaN Algeria 28.0339 1.6596 1/22/20 0 0 0
3 NaN Andorra 42.5063 1.5218 1/22/20 0 0 0
4 NaN Angola -11.2027 17.8739 1/22/20 0 0 0
In [4]:
df.drop(['Province/State'],axis=1,inplace=True)
df.head()
Out[4]:
Country/Region Lat Long Date Confirmed Deaths Recovered
0 Afghanistan 33.0000 65.0000 1/22/20 0 0 0
1 Albania 41.1533 20.1683 1/22/20 0 0 0
2 Algeria 28.0339 1.6596 1/22/20 0 0 0
3 Andorra 42.5063 1.5218 1/22/20 0 0 0
4 Angola -11.2027 17.8739 1/22/20 0 0 0
In [5]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 27984 entries, 0 to 27983
Data columns (total 7 columns):
Country/Region    27984 non-null object
Lat               27984 non-null float64
Long              27984 non-null float64
Date              27984 non-null object
Confirmed         27984 non-null int64
Deaths            27984 non-null int64
Recovered         27984 non-null int64
dtypes: float64(2), int64(3), object(2)
memory usage: 1.5+ MB
In [6]:
df['Date'] = pd.to_datetime(df['Date'])
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 27984 entries, 0 to 27983
Data columns (total 7 columns):
Country/Region    27984 non-null object
Lat               27984 non-null float64
Long              27984 non-null float64
Date              27984 non-null datetime64[ns]
Confirmed         27984 non-null int64
Deaths            27984 non-null int64
Recovered         27984 non-null int64
dtypes: datetime64[ns](1), float64(2), int64(3), object(1)
memory usage: 1.5+ MB
In [7]:
df.rename(columns={"Country/Region":"Country"},inplace=True)
df.head()
Out[7]:
Country Lat Long Date Confirmed Deaths Recovered
0 Afghanistan 33.0000 65.0000 2020-01-22 0 0 0
1 Albania 41.1533 20.1683 2020-01-22 0 0 0
2 Algeria 28.0339 1.6596 2020-01-22 0 0 0
3 Andorra 42.5063 1.5218 2020-01-22 0 0 0
4 Angola -11.2027 17.8739 2020-01-22 0 0 0
In [8]:
df['Active'] = df['Confirmed'] - df['Recovered'] - df['Deaths']
df.tail()
Out[8]:
Country Lat Long Date Confirmed Deaths Recovered Active
27979 Western Sahara 24.215500 -12.885800 2020-05-06 6 0 5 1
27980 Sao Tome and Principe 0.186360 6.613081 2020-05-06 174 3 4 167
27981 Yemen 15.552727 48.516388 2020-05-06 25 5 0 20
27982 Comoros -11.645500 43.333300 2020-05-06 8 1 0 7
27983 Tajikistan 38.861034 71.276093 2020-05-06 379 8 0 371

Latest Date Data

In [9]:
#### Latest Data
latest_date = df[df['Date']==df['Date'].max()]
latest_date.head()
Out[9]:
Country Lat Long Date Confirmed Deaths Recovered Active
27720 Afghanistan 33.0000 65.0000 2020-05-06 3392 104 458 2830
27721 Albania 41.1533 20.1683 2020-05-06 832 31 595 206
27722 Algeria 28.0339 1.6596 2020-05-06 4997 476 2197 2324
27723 Andorra 42.5063 1.5218 2020-05-06 751 46 521 184
27724 Angola -11.2027 17.8739 2020-05-06 36 2 11 23
In [10]:
world = latest_date.groupby('Country')['Confirmed','Deaths','Recovered','Active'].sum()
world = world.reset_index()
world.head()
Out[10]:
Country Confirmed Deaths Recovered Active
0 Afghanistan 3392 104 458 2830
1 Albania 832 31 595 206
2 Algeria 4997 476 2197 2324
3 Andorra 751 46 521 184
4 Angola 36 2 11 23
In [11]:
#### Plot on World Map (Active Cases)
figure = px.choropleth(world,locations='Country',locationmode='country names',
                       color='Active',range_color=[1,10000]
                       ,color_continuous_scale='ylorrd',title='World Map Plot')
figure.show()

World - Total Confirmed Cases

In [12]:
#### Plot WorldWide Confirmed Cases Over Data
World_Total_Confirmed = df.groupby('Date')['Confirmed'].sum().reset_index()
World_Total_Confirmed.tail()
Out[12]:
Date Confirmed
101 2020-05-02 3427337
102 2020-05-03 3506723
103 2020-05-04 3583049
104 2020-05-05 3662685
105 2020-05-06 3755335
In [13]:
plt.figure(figsize=(20,8))
plt.xticks(rotation=90,fontsize=8)
sns.pointplot(World_Total_Confirmed['Date'].dt.date,World_Total_Confirmed['Confirmed']
                  ,color='red')
plt.show()
In [14]:
plt.figure(figsize=(12,8))
plt.xticks(rotation=90,fontsize=7)
plt.plot(World_Total_Confirmed['Date'].dt.date,World_Total_Confirmed['Confirmed']
                  ,color='red')
plt.scatter(World_Total_Confirmed['Date'].dt.date,World_Total_Confirmed['Confirmed']
                  ,color='blue')
plt.show()

C:\Users\VAIO\Anaconda3\lib\site-packages\pandas\plotting\_matplotlib\converter.py:103: FutureWarning:

Using an implicitly registered datetime converter for a matplotlib plotting method. The converter was registered by pandas on import. Future versions of pandas will require you to explicitly register matplotlib converters.

To register the converters:
	>>> from pandas.plotting import register_matplotlib_converters
	>>> register_matplotlib_converters()

Active - Top 20 Countries

In [15]:
##### Top 20 Countries having most number of Active Cases
Active_Top_Countries = latest_date.groupby('Country')['Active'].sum().sort_values(ascending=False)
Active_Top_Countries = Active_Top_Countries.reset_index()
Active_Top_Countries.head()
Out[15]:
Country Active
0 US 965262
1 United Kingdom 171275
2 Russia 143065
3 France 94333
4 Italy 91528
In [16]:
Active_Top_20_Countries = Active_Top_Countries.head(20)
In [17]:
plt.figure(figsize=(8,8))
sns.barplot(Active_Top_20_Countries['Active'],Active_Top_20_Countries['Country'], label = "Active Cases")
plt.xlabel("Active",  fontdict = {'fontsize' : 15})
plt.ylabel("Country",  fontdict = {'fontsize' : 15})
plt.title("Top 20 countries with Active COVID-19 cases",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()

Recovered - Top 20 Countries

In [18]:
##### Top 20 Countries having most number of Recovered Cases
Recovered_Top_Countries = latest_date.groupby('Country')['Recovered'].sum().sort_values(ascending=False)
Recovered_Top_Countries= Recovered_Top_Countries.reset_index()
Recovered_Top_20_Countries = Recovered_Top_Countries.head(20)
In [19]:
plt.figure(figsize=(8,8))
sns.barplot(Recovered_Top_20_Countries['Country'],Recovered_Top_20_Countries['Recovered'], label = "Recovered")
plt.tick_params(axis="x",labelrotation=90, labelsize = 12)
plt.xlabel("Recovered",  fontdict = {'fontsize' : 15})
plt.ylabel("Country",  fontdict = {'fontsize' : 15})
plt.title("Top 20 countries with Recovered COVID-19 cases",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()
plt.show()

Confirmed - Top 20 Countries

In [20]:
##### Top 20 Countries having most number of Confirmed Cases
Confirmed_Top_Countries  = latest_date.groupby('Country')['Confirmed'].sum().sort_values(ascending=False)
Confirmed_Top_Countries= Confirmed_Top_Countries.reset_index()
Confirmed_Top_20_Countries = Confirmed_Top_Countries.head(20)
In [21]:
plt.figure(figsize=(8,8))
sns.barplot(Confirmed_Top_20_Countries['Country'],Confirmed_Top_20_Countries['Confirmed'], label = "Confirmed")
plt.tick_params(axis="x",labelrotation=90, labelsize = 12)
plt.xlabel("Confirmed",  fontdict = {'fontsize' : 15})
plt.ylabel("Country",  fontdict = {'fontsize' : 15})
plt.title("Top 20 countries with Confirmed COVID-19 cases",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()

Confirmed/Recovered - Top 20 Countries

In [22]:
plt.figure(figsize=(8,8))
sns.barplot(Confirmed_Top_20_Countries['Confirmed'],Confirmed_Top_20_Countries['Country'],color='red', label = "Confirmed")
sns.barplot(Recovered_Top_20_Countries['Recovered'],Recovered_Top_20_Countries['Country'],color='green', label = "Recovered")
plt.xlabel("Scale",  fontdict = {'fontsize' : 15})
plt.ylabel("Country",  fontdict = {'fontsize' : 15})
plt.title("Top 20 countries with Confirmed and Recovered COVID-19 cases",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()

India - COVID-19 Analysis

In [23]:
#### Make some dataframes for some countries
India = df[df['Country']=='India']
India = India.groupby('Date')['Recovered','Deaths','Active','Confirmed'].sum().reset_index()
India.tail()
Out[23]:
Date Recovered Deaths Active Confirmed
101 2020-05-02 10819 1323 27557 39699
102 2020-05-03 11775 1391 29339 42505
103 2020-05-04 12847 1566 32024 46437
104 2020-05-05 14142 1693 33565 49400
105 2020-05-06 15331 1785 35871 52987
In [24]:
plt.figure(figsize=(5,5))
plt.plot(India['Date'],India['Confirmed'],color='blue', label = "Confirmed")
plt.plot(India['Date'],India['Active'],color='orange', label = "Active")
plt.plot(India['Date'],India['Deaths'],color='red', label = "Deaths")
plt.plot(India['Date'],India['Recovered'],color='green', label = "Recovered")
plt.tick_params(axis="x",labelrotation=90, labelsize = 12)
plt.xlabel("Scale",  fontdict = {'fontsize' : 15})
plt.ylabel("Country",  fontdict = {'fontsize' : 15})
plt.title("India Covid-19 analysis",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()

Recovered Comparision - India, Italy, China & US

In [25]:
### Italy , US, China & India
Italy = df[df['Country']=='Italy']
Italy = Italy.groupby('Date')['Recovered','Deaths','Active','Confirmed'].sum().reset_index()
US = df[df['Country']=='US']
US = US.groupby('Date')['Recovered','Deaths','Active','Confirmed'].sum().reset_index()
China = df[df['Country']=='China']
China = China.groupby('Date')['Recovered','Deaths','Active','Confirmed'].sum().reset_index()
In [26]:
plt.figure(figsize=(12,8))
plt.plot(China.index,China['Recovered'],color='Red',label='China')
plt.plot(India.index,India['Recovered'],color='Green',label='India')
plt.plot(US.index,US['Recovered'],color='Black',label='US')
plt.plot(Italy.index,Italy['Recovered'],color='Blue',label='Italy')
plt.scatter(China.index,China['Recovered'],color='Red')
plt.scatter(India.index,India['Recovered'],color='Green')
plt.scatter(US.index,US['Recovered'],color='Black')
plt.scatter(Italy.index,Italy['Recovered'],color='Blue')
plt.legend(loc=2)
plt.show()

Prediction and Forecasting

In [27]:
#### Library - Fbprophet
#### Created by Facebook Company for time series analysis
#### The two column should be named as ds(date) and y(data)
#### Where to Apply time series -
#### 1. When the data is not constant
#### 2. When the data is not following any function
In [28]:
from fbprophet import Prophet
In [29]:
df.head()
Out[29]:
Country Lat Long Date Confirmed Deaths Recovered Active
0 Afghanistan 33.0000 65.0000 2020-01-22 0 0 0 0
1 Albania 41.1533 20.1683 2020-01-22 0 0 0 0
2 Algeria 28.0339 1.6596 2020-01-22 0 0 0 0
3 Andorra 42.5063 1.5218 2020-01-22 0 0 0 0
4 Angola -11.2027 17.8739 2020-01-22 0 0 0 0
In [30]:
confirmed = df.groupby('Date')['Confirmed'].sum().reset_index()
confirmed.head()
Out[30]:
Date Confirmed
0 2020-01-22 555
1 2020-01-23 654
2 2020-01-24 941
3 2020-01-25 1434
4 2020-01-26 2118
In [31]:
confirmed.columns = ['ds','y']
In [32]:
#### Build The Model
model = Prophet(interval_width=0.95)
In [33]:
### Train the model
model.fit(confirmed)

INFO:fbprophet:Disabling yearly seasonality. Run prophet with yearly_seasonality=True to override this.
INFO:fbprophet:Disabling daily seasonality. Run prophet with daily_seasonality=True to override this.
Out[33]:
<fbprophet.forecaster.Prophet at 0x26ad32a2c08>
In [34]:
future_dates = model.make_future_dataframe(periods=7)
future_dates.tail(10)
Out[34]:
ds
103 2020-05-04
104 2020-05-05
105 2020-05-06
106 2020-05-07
107 2020-05-08
108 2020-05-09
109 2020-05-10
110 2020-05-11
111 2020-05-12
112 2020-05-13
In [35]:
forecast = model.predict(future_dates)
In [36]:
forecast[['ds','yhat','yhat_lower','yhat_upper']].tail(10)
Out[36]:
ds yhat yhat_lower yhat_upper
103 2020-05-04 3.584437e+06 3.575971e+06 3.592767e+06
104 2020-05-05 3.662043e+06 3.653052e+06 3.670132e+06
105 2020-05-06 3.742727e+06 3.734510e+06 3.751179e+06
106 2020-05-07 3.824092e+06 3.815124e+06 3.834171e+06
107 2020-05-08 3.907160e+06 3.895138e+06 3.918149e+06
108 2020-05-09 3.987840e+06 3.968499e+06 4.007947e+06
109 2020-05-10 4.067524e+06 4.036644e+06 4.095628e+06
110 2020-05-11 4.142802e+06 4.102453e+06 4.180020e+06
111 2020-05-12 4.220407e+06 4.168035e+06 4.268686e+06
112 2020-05-13 4.301092e+06 4.235185e+06 4.361759e+06
In [37]:
confirmed.tail()
Out[37]:
ds y
101 2020-05-02 3427337
102 2020-05-03 3506723
103 2020-05-04 3583049
104 2020-05-05 3662685
105 2020-05-06 3755335
In [38]:
confirmed_plot = model.plot(forecast)
In [39]:
##### Top 20 Countries having most number of Active Cases
# top_data = latest_date.groupby('Country')['Confirmed','Recovered'].sum()
# top_data = top_data.sort_values('Confirmed',ascending=False)
# top_data = top_data.reset_index()
# top_data.head()
# top_20 = top_data.head(20)
In [40]:
# plt.figure(figsize=(8,8))
# sns.barplot(top_20['Confirmed'],top_20['Country'],color='red')
# sns.barplot(top_20['Recovered'],top_20['Country'],color='green')
# plt.show()

INDIA - Covid Prediction

In [41]:
India_Prediction = df[df['Country']=='India']
India_Prediction.tail()
Out[41]:
Country Lat Long Date Confirmed Deaths Recovered Active
26795 India 21.0 78.0 2020-05-02 39699 1323 10819 27557
27059 India 21.0 78.0 2020-05-03 42505 1391 11775 29339
27323 India 21.0 78.0 2020-05-04 46437 1566 12847 32024
27587 India 21.0 78.0 2020-05-05 49400 1693 14142 33565
27851 India 21.0 78.0 2020-05-06 52987 1785 15331 35871
In [42]:
grouped_India_Prediction = India_Prediction.groupby('Date')['Confirmed'].sum().reset_index()
grouped_India_Prediction.rename(columns={'Date':'ds','Confirmed':'y'},inplace=True)
grouped_India_Prediction.tail()
Out[42]:
ds y
101 2020-05-02 39699
102 2020-05-03 42505
103 2020-05-04 46437
104 2020-05-05 49400
105 2020-05-06 52987
In [43]:
#### Build The Model
india_model = Prophet(interval_width=0.95)
In [44]:
### Train the model
india_model.fit(grouped_India_Prediction)

INFO:fbprophet:Disabling yearly seasonality. Run prophet with yearly_seasonality=True to override this.
INFO:fbprophet:Disabling daily seasonality. Run prophet with daily_seasonality=True to override this.
Out[44]:
<fbprophet.forecaster.Prophet at 0x26ad38c22c8>
In [45]:
future_prediction_dates = model.make_future_dataframe(periods=7)
future_prediction_dates.tail(7)
Out[45]:
ds
106 2020-05-07
107 2020-05-08
108 2020-05-09
109 2020-05-10
110 2020-05-11
111 2020-05-12
112 2020-05-13
In [46]:
prediction = india_model.predict(future_prediction_dates)
prediction.head(3)
Out[46]:
ds trend yhat_lower yhat_upper trend_lower trend_upper additive_terms additive_terms_lower additive_terms_upper weekly weekly_lower weekly_upper multiplicative_terms multiplicative_terms_lower multiplicative_terms_upper yhat
0 2020-01-22 -114.608959 -1812.995443 2274.271553 -114.608959 -114.608959 251.411646 251.411646 251.411646 251.411646 251.411646 251.411646 0.0 0.0 0.0 136.802687
1 2020-01-23 -109.756882 -2477.435452 2046.006106 -109.756882 -109.756882 -160.604681 -160.604681 -160.604681 -160.604681 -160.604681 -160.604681 0.0 0.0 0.0 -270.361563
2 2020-01-24 -104.904806 -2205.618824 1882.179924 -104.904806 -104.904806 -183.668113 -183.668113 -183.668113 -183.668113 -183.668113 -183.668113 0.0 0.0 0.0 -288.572919
In [47]:
prediction_result = prediction[['ds','yhat','yhat_lower','yhat_upper']]
prediction_result.rename(columns={'yhat':'predicted','yhat_lower':'lower_limit','yhat_upper':'upper_limit'}, inplace=True)
prediction_result.tail(8)

C:\Users\VAIO\Anaconda3\lib\site-packages\pandas\core\frame.py:4223: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
Out[47]:
ds predicted lower_limit upper_limit
105 2020-05-06 46955.527987 44889.099426 49107.128605
106 2020-05-07 48223.755251 46239.294201 50302.405435
107 2020-05-08 49880.935409 47734.298883 51894.250027
108 2020-05-09 51621.516891 49611.538592 53821.375868
109 2020-05-10 53362.234670 51248.314580 55465.249111
110 2020-05-11 55200.396047 52912.347405 57318.937536
111 2020-05-12 56969.563530 54696.073200 59175.735401
112 2020-05-13 58717.233119 56446.694700 61052.085453

Trend chart for month April till mentioned forcast day

In [48]:
prediction_month = prediction_result.iloc[-40:,:]

#import datetime
plt.figure(figsize=(15,5))
plt.plot(prediction_month['ds'],prediction_month['predicted'],color='blue', label = "Predicted")
plt.tick_params(axis="x",labelrotation=90, labelsize = 12)
#plt.xlim([datetime.date(2020,4,1), datetime.date(2020, 4, 15)])
plt.xlabel("Date",  fontdict = {'fontsize' : 15})
plt.ylabel("Count",  fontdict = {'fontsize' : 15})
plt.title("India Covid-19 predictions for April Month",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()
In [49]:
prediction_dates = prediction_result.iloc[-7:,:]

plt.figure(figsize=(10,5))
plt.plot(prediction_dates['ds'],prediction_dates['predicted'],color='red', label = "Predicted")
plt.tick_params(axis="x",labelrotation=90, labelsize = 12)
plt.xlabel("Date",  fontdict = {'fontsize' : 15})
plt.ylabel("Count",  fontdict = {'fontsize' : 15})
plt.title("India Covid-19 future 7 days predictions",  fontdict = {'fontsize' : 15})
plt.legend()
plt.show()
In [ ]: